This project suggests a quick, easy, and non-invasive method for detecting a person’s blood group using fingerprint images and a deep learning technique. The goal of the project is to improve blood group classification in medical situations involving organ transplants, blood donations, and critical care. The method involves preprocessing approaches, data augmentation to address class imbalance, and model training with 30 epochs for evaluation constitute every part of the strategy. The technique’s efficiency is demonstrated by the results, which show an amazing best accuracy. Users can contribute fingerprint images for blood group prediction using an intuitive Flask-based web application that incorporates the best-performing model. Based on the study’s findings, blood group detection can be made simpler, which is a huge improvement in healthcare.